Reconfiguration in network of embedded systems: Challenges and adaptive tracking case study

  • Authors:

  • Soheil Ghiasi;Ani Nahapetian;Hyun J. Moon;Majid Sarrafzadeh

  • Affiliations:

  • Computer Science Department, University of California (UCLA), Los Angeles, CA 90095, USA (Corresponding author. Tel.: +1 310 794 5616/ Fax: +1 310 794 5056/ E-mail: soheil@cs.ucla.edu);Computer Science Department, University of California (UCLA), Los Angeles, CA 90095, USA;Computer Science Department, University of California (UCLA), Los Angeles, CA 90095, USA;Computer Science Department, University of California (UCLA), Los Angeles, CA 90095, USA

  • Venue:
  • Journal of Embedded Computing - Real-Time and Embedded Computing Systems
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Many applications utilize deeply embedded sensors and actuators that are tightly coupled with the physical environment in order to perform their functionality. Sensor, actuators and embedded computation resources used for implementing such systems usually exhibit regular local configurations, while the global structure of the subsystems is either not fixed a priori and can change at runtime or is not known. Examples include systems that use many randomly distributed sensing boards, each one having a fixed structure of computation resources and sensing devices, to autonomously detect events and take proper actions. This paper discusses the requirements of the aforementioned systems, their advantages and the issues involved in developing them. Specifically we focus on dynamic adaptation of the system as a particular feature of such systems. This feature is discussed in depth in a collaborative and dynamically adaptive object tracking system that has been built in our lab as the experimental framework of this study. We exploit reconfigurable hardware devices embedded in a number of networked cameras in order to achieve our goal. We justify the need for dynamic adaptation of the system through scenarios and applications. Experimental results on a set of scenes advocate the fact that our system works effectively for different scenario of events through reconfiguration. Comparing results with non-adaptive implementations verify the fact that our approach improves system's robustness to scene variations and outperforms the traditional implementations.